22 billion devices will be connected to the Internet of Things (IoT) by the end of this year, and that number is expected to surge to almost 39 billion by the middle of the decade.

When the Wintel amigos – Microsoft’s Bill Gates and Andy Groves from Intel – were planning, then executing, world domination in the IT sector at the end of the last century, they famously imagined a computer on every desktop.

Gates later admitted that what he didn’t see coming was computers in every consumer device, in every jacket pocket, on every item in the fridge, and indeed in anything that might conceivably find cause to send a data point to a network.

Worldwide spending on the IoT exceeded a trillion dollars a year in 2020 according to industry analyst IDC, and growth is expected to exceed double digits this year.

In 2015 McKinsey and Company estimated that the IoT had the potential to generate $US4–11 trillion in annual economic value by 2025. Since then, they’ve upgraded their estimate to $5.5–12.6 trillion in annual economic value — which includes the value captured by consumers and customers of IoT products and services.

Whilst McKinsey remain bullish overall, they do note the “IoT has faced headwinds related to change management, cost, talent, and cybersecurity, particularly in enterprises.”

That said, as more investment dollars flood into business models built around the convergence of the digital and physical worlds, we see the proliferation of the connected thing as a fundamental trend.

About those things? 

In the simplest reading, the Internet of Things refers to the ways in which devices that are not thought of as computers, like home appliances, cars, clothing — things, as it were — are connected in the way computers have traditionally been.

Obviously things don’t have much computing power on board, but that doesn’t matter. John Gage’s immortal aphorism “the network is the computer” truly comes into its own with the IoT. The devices themselves don’t need a whole lot of power, because the real magic happens on IoT platforms — where the data is collected, analysed, and processed. It’s an important point, since a great many of those things are simply sensors — including those which exist only to remind their owners of their existence — yet even that one simple data point is incredibly valuable in a world of data analytics, as anyone who works in predictive maintenance can attest.

The applications of the IoT are immense. Everything from the “smart home” — with appliances that talk to each other so your coffee pot knows when your morning alarm goes off and your lights adjust themselves according to the time of day — to the “industrial IoT” with mines and factories that essentially run themselves using data collected from millions of devices. Imagine supply chains that could automatically adjust themselves according to fluctuations in supply and demand, or even in response to natural disasters. 

Self-driving cars are an obvious place where the ability to collect an enormous amount of data, process it and communicate it quickly is essential. But applications for the IoT can be found in almost any industry — including retail, medicine and even law enforcement. For insight into how companies across SEA and ANZ are integrating IoT technology, in our second article we’ve spotlighted three companies we know well which are doing just that.

Gartner’s figures indicate that right now the largest IoT application is surveillance devices (both in the home and in public spaces). But by 2025 that will be overtaken by IoT automotive devices, as self-driving vehicles move closer to reality.

Enabling the IoT 

Back in 1999, the biggest obstacle standing in the way of the IoT was the internet itself. Every computer on the internet, every printer, every device (indeed, every thing) needs a unique IP address, and the system for allocating addresses — IP version 4 — only had capacity for around 4.6 billion unique devices. That’s not nearly enough. IPv6, which was only ratified in 2017, allows for 2^128 unique addresses — that’s a very very big number. The same year IPv6 was ratified, connected devices outnumbered people in the world for the first time.

Then of course there’s the size of computing tech. Chips and batteries are getting smaller all the time (it has always been thus) so naturally the amount of computing power you can fit into a smaller package is always increasing. Power usage remains an issue, but not to the extent it once was.

The next major enabling technology is communication. It’s no use having a connected device unless it can connect. Bluetooth, Wi-Fi and satellite communications, and to an extent 4G, have been the mainstay of IoT so far — but 5G will represent a step-change in how many devices can be connected, and the speed at which they can exchange data. The key advantage of 5G is that it enables incredible density — up to a million devices per square kilometre. 4G simply cannot handle that many connections so close together, which has slowed the expansion of IoT use in real world environments — particularly in industrial settings.

Finally, the sheer volume of data to be processed represents a major hurdle for the IoT. And that is where the growth of massive data centres and cloud computing has come into its own. Few companies have the resources to maintain the kind of data processing that a large IoT implementation calls into play. But by outsourcing the analytics to cloud providers, the costs become more manageable. And the existence of such huge pools of data has enabled advancements in machine learning and AI, which had been progressing quite slowly in the era before everything was connected.

The next development to watch for, though, is “edge” computing. As the “cloud” model matures, more companies are recognising the limitations of centralised data processing — even with 5G. Increasingly, we expect to see IoT devices with more smarts built into them, so that processing can happen closer to where the devices are collecting the data.

The elephant in the room

So now that everything’s connected and every device is a smart device, the world is a safer place, right?

Wrong.

There will always be bad actors looking to cause mischief on computing networks, and the IoT offers an enormous number of points of vulnerability. Security professionals who once needed to lock down a few hundred devices to keep their organisations safe now have to think in terms of millions of devices.

What’s more, these devices are in many places connected to the real world. It’s one thing if a breach causes a company to lose data, quite another if it means the temperature sensors in a power plant stop working — or the brakes on a car. The potential for malicious action is massive.

The security industry is aware of the issue, but security solutions have not kept pace with the growth of IoT – and certainly not with the number of device manufacturers hopping on the bandwagon.

Then there’s the issue of data privacy. Not long ago, users were, on the whole, happy to hand over personal information about themselves provided it gave them a better experience whilst using the web. However, since scandals like Cambridge Analytica, there has been a growing awareness of the ways in which personal data can be misused. Users are becoming more protective of their privacy and reluctant to hand over data. 

That means the floods of data that massive analytic and AI systems need in order to do their magic are under threat, representing a problem for the IoT. Trust must be won back.

The Internet of Things is doubtless where the connected world is heading, but until it can prove that it’s safe for users as well as the companies that rely on its data, it risks falling short of its potential.